سال انتشار: ۱۳۹۱

محل انتشار: پانزدهمین کنفرانس دانشجویی مهندسی برق ایران

تعداد صفحات: ۵

نویسنده(ها):

Milad Ghahari Bidgoli – Faculty Of Engineering, Islamic Azad University – South Tehran Branch
Ali Broumandnia – Faculty Of Engineering, Islamic Azad University – South Tehran Branch
Mahdi Esmaeili – Faculty Of Engineering, Islamic Azad University – Kashan

چکیده:

In recent decades, research in motion mining and correct tracking without error in the proper time has been high importance. This paper presents a new model for motion miningusing a combination of data tracker-level strategies and then we utilize an additional layer as a tracker-level fusion to achieve the best results at the surface. In fact in this model, Fusion at tracker level models such as a black box that its output will be achieved by combining different algorithms and images features. The most important advantages of this model relative to other existing models are acceptable speed and accuracy for tracking in the higher layers